Design and Implementation - Segmentation of Hand Bone for Bone Age Assessment - page 52

(MBOBHE). The second purpose is realized by adopting the anisotropic diffusion
technique which is conventionally adopted in ultrasound image noise reduction
process.
The first aim of equalizing the brightness is to standardize the image and this in
turn increases the robustness of the proposed procedure to various hand bone radi-
ographs from different X-ray machine with different range of intensity distribu-
tion due to environmental factor or the devices internal factor. The second aim of
equalizing the brightness is to enhance the contrast of the input image to ease the
detection of the pertinent features. Both aims prepare the input image to improve
the result of subsequent procedures. The technique adopted is histogram equaliza-
tion with some modifications and improvements. Firstly, the review on histogram
equalization technique is conducted to assure that the current histogram equaliza-
tion methods are unable to cope with the inherent problems of hand bone radio-
graph contrast enhancement. Then, modified histogram equalization is proposed
and designed to optimize the visual effect of resultant equalized image for multiple
objectives. The details of the designed histogram equalization method are dis-
cussed in next section.
3.3.1 The Proposed MBOBHE
The contrast of bone texture in ossification sites is of critical importance in the
bone age assessment especially for TW3 method. An optimum contrast adjusted
ossification site image can ease the subsequent procedures in computer-aided skel-
etal age scoring system. The first advantage is that the outline of the bone structure
can be perceived clearly. Second advantage is devoted to ossification development
stage evaluation procedures by making both the structure and texture of ossification
bones to be well-defined and the important features are visually and computation-
ally distinguishable. Hence, a good contrast as pre-processing in computer-aided
skeletal age scoring system is prominent for accurate determination of bone age.
Therefore, the objective is to find an optimal solution to enhance the contrast,
the brightness and detail preservation of output image simultaneously. A new
histogram equalization method is proposed, named Multipurpose Beta Optimal
Bi-histogram Equalization (MBOBHE). This method disagrees with traditional
improvements of histogram equalization that tend to be optimized by minimizing
or maximizing only one of the properties such as contrast, brightness preservation
and detail preservation. Instead, a histogram equalization that capable of optimiz-
ing all the properties simultaneously by using regularization function rather than
by minimization or maximization any single objective such as brightness preserva-
tion, contrast enhancement or detail preservation is proposed. The motivation of
the design is due to the fact which we will prove qualitatively in this topic later,
that a visually good contrast image should be optimized in terms of holistic perfor-
mance of all properties. Holistic performance that takes all properties into account
can produce more natural output image.